scilab ebuilds only checks for g77 fortran compiler, while gcc4 calls it gfortran. trying to bypass this by symlink gives error during configure. please check
Created attachment 82746 [details] ebuild to utilize fortran.eclass for gfortran checking Try this...it uses fortran.eclass to check for gfortran. I don't have gcc-4 to try it out though...
the ebuild keeps 2.7 version, instead of 3.1.1 however i'm testing..wait some hours..
i was too optimistic emerge scilab Calculating dependencies... done! >>> Emerging (1 of 1) sci-mathematics/scilab-3.1.1-r3 to / >>> checksums files ;-) scilab-2.7-r3.ebuild >>> checksums files ;-) scilab-3.1.1-r3.ebuild >>> checksums files ;-) files/digest-scilab-3.1.1-r3 >>> checksums files ;-) files/scilab-2.7-configure.patch >>> checksums files ;-) files/gzvt2-add-gtk2-flags.patch >>> checksums files ;-) files/digest-scilab-2.7-r3 >>> checksums src_uri ;-) scilab-3.1.1-src.tar.gz >>> checksums src_uri ;-) patch_browsehelp.tar.gz which: no scilab in (/sbin:/usr/sbin:/usr/lib/portage/bin:/bin:/usr/bin:/opt/bin:/usr/athena/sbin:/usr/i686-pc-linux-gnu/gcc-bin/4.1.0:/opt/intel/compiler90/bin:/opt/intel/fortran90/bin:/opt/xmingw/bin:/opt/stuffit/bin:/opt/sun-jdk-1.5.0.06/bin:/opt/sun-jdk-1.5.0.06/jre/bin:/usr/bin:/usr/kde/3.5/sbin:/usr/kde/3.5/bin:/usr/qt/3/bin:/opt/vmware/workstation/bin:/usr/bin/scripts:/sbin:/usr/sbin/:/usr/bin:/opt/bin:/bin:/usr/X11R6/bin:/usr/share/karamba/bin) which: no g77 in (/sbin:/usr/sbin:/usr/lib/portage/bin:/bin:/usr/bin:/opt/bin:/usr/athena/sbin:/usr/i686-pc-linux-gnu/gcc-bin/4.1.0:/opt/intel/compiler90/bin:/opt/intel/fortran90/bin:/opt/xmingw/bin:/opt/stuffit/bin:/opt/sun-jdk-1.5.0.06/bin:/opt/sun-jdk-1.5.0.06/jre/bin:/usr/bin:/usr/kde/3.5/sbin:/usr/kde/3.5/bin:/usr/qt/3/bin:/opt/vmware/workstation/bin:/usr/bin/scripts:/sbin:/usr/sbin/:/usr/bin:/opt/bin:/bin:/usr/X11R6/bin:/usr/share/karamba/bin) * No fortran compiler found on the system! * Please add fortran to your USE flags and reemerge gcc! !!! ERROR: sci-mathematics/scilab-3.1.1-r3 failed. Call stack: ebuild.sh, line 1581: Called dyn_setup ebuild.sh, line 686: Called pkg_setup !!! (no error message) !!! If you need support, post the topmost build error, and the call stack if relevant.
Get rid of this in pkg_setup: if ! use ifc && [ -z `which g77` ]; then #if ifc is defined then the dep was already checked eerror "No fortran compiler found on the system!" eerror "Please add fortran to your USE flags and reemerge gcc!" die fi Also the ifc in IUSE and in the dependencies.
but...have you tested it? guess not :) the configure doesn't check for gfortran...should be updated...but...however... checking for javac... yes checking for javac... /opt/sun-jdk-1.5.0.06/bin/javac checking for leading and/or trailing underscores... nm: 'pipof.o': No such file nm: 'pipof.o': No such file nm: 'pipof.o': No such file no no checking for use of sharpsign in CPP... no checking for main in -lm... no checking for exp10... no checking for getwd... no checking for sleep... no checking for strerror... no checking for usleep... no checking for finite... no checking limits.h usability... no checking limits.h presence... yes configure: WARNING: limits.h: present but cannot be compiled configure: WARNING: limits.h: check for missing prerequisite headers? configure: WARNING: limits.h: proceeding with the preprocessor's result configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-autoconf@gnu.org. ## configure: WARNING: ## ------------------------------------ ## checking for limits.h... yes checking values.h usability... no checking values.h presence... yes configure: WARNING: values.h: present but cannot be compiled configure: WARNING: values.h: check for missing prerequisite headers? configure: WARNING: values.h: proceeding with the preprocessor's result configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-autoconf@gnu.org. ## configure: WARNING: ## ------------------------------------ ## checking for values.h... yes checking for main in -lncurses... no checking for main in -lcurses... no checking for main in -ltermcap... no checking for main in -ltermlib... no checking for PVM architecture... LINUX checking for main in -ldl... no checking for header file tcl.h... found in /usr/include checking if tcl is version 8.4 or later... cc1: error: unrecognized command line option "-fwritable-strings" (can't happen) yes checking for tcl library tcl.... found /usr/lib/libtcl.so using -L/usr/lib -ltcl checking for Tcl_DoOneEvent in -ltcl... no configure: error: header file tcl.h has been found for 8.4* but no corresponding tcl library !!! Please attach the following file when filing a report to bugs.gentoo.org: !!! /var/tmp/portage/scilab-3.1.1-r3/work/scilab-3.1.1/config.log !!! ERROR: sci-mathematics/scilab-3.1.1-r3 failed. Call stack: ebuild.sh, line 1565: Called dyn_compile ebuild.sh, line 974: Called src_compile scilab-3.1.1-r3.ebuild, line 64: Called econf '--with-gtk' !!! econf failed !!! If you need support, post the topmost build error, and the call stack if relevant. the log file complaing about "-fwritable-strings" switch.
(In reply to comment #5) > but...have you tested it? guess not :) > > the configure doesn't check for gfortran...should be updated...but...however... No, and I don't have time to. I'm trying to point you in the right direction. The configure.ac file may need an edit to remove that -fwritable-strings (which hopefully the code no longer uses, else you'll likely get a compile error later on), and then you can inherit autotools.eclass and eautoreconf.
cosidering that scilab 3 is pretty old and that's scilab 4 with a long changelog (http://www.scilab.org/download/features/scilab40.php) it should be better to bump to 4 and test it. i guess 3 series it unmainteined upstream too, so they won't help wioth gcc 4.x bump to 4 and i'll be happy to test
i close this bug, it's stupid to work on version 3 while in bugzilla that's activity against version 4
